takahirom/arbigent
An AI agent testing framework that uses LLMs to autonomously perform UI testing across Android, iOS, and Web platforms.

Arbigent (Arbiter-Agent) is an AI agent testing framework designed to automate UI testing across Android, iOS, and Web applications. It breaks down complex testing scenarios into smaller, dependent tasks that AI agents can execute predictably, addressing challenges like UI changes, dynamic content, and unexpected dialogs that typically break traditional tests. The framework supports multiple AI providers (Gemini, OpenAI) and provides both a UI and code interface for test configuration and execution.